from that position, and is serving as Professor and

Previous studies of the gender pay gap in the United States have found a narrowing of that gap in the past two decades, but have differed substantially in the reported magnitude of that change. This study calculates gender earnings differences using Panel Study of Income Dy-namics data for 1968-89, a longer time series than has been examined by any similar study. It also employs improved statistical techniques, notably an adjustment for selection bias and a recently recommended modification of the most commonly used decomposition technique. The author finds a significant decline both in the portion of the differential due to observed differences between men's and women's characteristics and in the "unexplained " component, which is often attributed to discrimination. The results of one analysis show that over the 1968-89 period the gender pay gap narrowed to as little as 47%. T he magnitude of the male-female earnings gap and changes in this dif-ferential over time have been the focus of much attention among researchers and policy-makers. Of the several studies that have examined the male-female earnings gap, however, only a few have tracked the change in that gap over a sustained period of time. Those few studies agreed that th